Quantum computing has been a hot topic amongst eager tech-savvy entrepreneurs and academics alike for decades. In recent years advances in nanofabrication techniques have made the concept of quantum computing a reality.
However, it remains to be seen which approach will yield the most commercially viable quantum computer – will it be superconducting systems, ion-traps, neutral atoms or something else?
Why is quantum computing desirable?
The first quantum computers already exist and some are even available for purchase or use online – so why has it not yet taken off?
All quantum computers can function only if they have the ability to create, maintain, and control qubits. What your computer uses as a qubit defines your modality of quantum computing, and each type of qubit has its own benefits and drawbacks.
Of all the problems that each modality faces, the chief issue holding them back is scalability. Great strides have been made in the maximum number of qubits that modern quantum computers can sustain, but the fact remains that we will not see the true benefits of quantum computing until large numbers of qubits can be produced, maintained, and interacted with.
How do you make a qubit?
The fundamental requirement of a qubit is that it is an object that behaves quantum mechanically. It must be capable of being described as having a quantum mechanical state, and that state must be predictable/programmable.
For it to be a useful qubit and not just a nanoscale ornament for one’s quantum trophy cabinet, you must also be able to interact with the qubit without forcing it out of the quantum regime. This means you must be able to prepare it in a state, let the state evolve, and then read out the state.
The last big requirement is that you must be able to entangle the qubit with another qubit at will.
A promising starting point for a qubit then is anything that behaves quantum mechanically and can be easily controlled.
Superconducting Qubits
This is the solution that had an early start due to the quantum nature of superconductivity. In brief, at low temperatures some materials allow electrons to form ‘Cooper pairs’, which can travel through a material without experiencing resistance. Since these Cooper pairs are achievable simply by making a particular material cold, it is relatively simple to create a circuit that contains these pairs, which are already in a quantum mechanical state.
However, since such circuits are restricted in geometry to the shapes that can be easily fabricated, the lifetime of a state on such circuits (the coherence time) suffers and can be expected around the microsecond to millisecond range. This however is not the biggest obstacle for this modality – that is temperature.
There is a non-trivial obstacle in keeping superconducting qubits at a low enough temperature for them to behave quantum mechanically, and this is complicated by the fact that a lot of the electronics required for communication with the qubits need to access the system. In many implementations, there is a strong thermal coupling between these electronics and the qubits, so the energy and equipment requirement for this approach can be prohibitive.
Trapped Ion Qubits
An approach that has been fast gathering pace in the world of quantum computing is using individual atoms as qubits.
Ion traps have been known for some time, but it has not been until recently, with the advent of the quantum charge coupled device, that computation with ions as qubits has become readily accessible. Coherence times of ions is superior to that of superconducting qubits, with some groups reporting coherence times of over an hour at low temperatures.
Unlike superconducting qubits, these qubits do not exist if the quantum computer is ‘turned off’. A complex process of vaporising elements and ionising the resultant atomic cloud produces ions of the various species selected, which are then trapped in electric fields and cooled using lasers.
The cooled ions are then shuttled to the appropriate part of the computer where they are trapped and pumped by various lasers to perform the quantum computation.
Here the problem again is scalability: an object of large scale quantum computing is to create large systems of entangled qubits. Whereas in superconducting systems one can just ‘print’ the qubit where one wants it, it is far harder to arrange ions near one another in a large enough array of qubits.
Entangling trapped ions is also difficult, because they naturally repel one another, so the process of entangling two ions often involves the use of photons travelling between the two ions, rather than direct interaction of the ions themselves.
However, progress is being made in that recent research has demonstrated that qubits can be moved from one quantum charge coupled device to another without losing coherence. This opens up the possibility of arranging many such devices next to one another to create a sufficiently large computer.
Neutral Atom Qubits
An even more recent entrant to the race is qubits made from neutral atoms. This is very similar in concept to the trapped ion approach, except that the atoms are not ionised. Instead, they are trapped using magnetic fields and lasers (as opposed to the electric fields of the ion trapping system).
The strength of magnetic field required is less than that of the electric fields required in trapped ions. Furthermore, because the qubits do not naturally repel one another they can simply be placed nearby and then excited with lasers such that their electron orbitals overlap, allowing entanglement. Each atom is held by its own tuned laser - so-called ‘optical tweezers’ - which will interact only with that particular species of atom.
Unlike other systems, the atoms then can be quickly and easily moved to another location in the computer by the optical tweezers, making this modality more flexible than other modalities.
This new approach also promises high coherence times, and a lack of requirement for complex cryogenic equipment.
The main drawback of neutral atoms is simply that the technology is still new and has not had time to mature as the other two mainline modalities have.
It is probably too early to say which modality will win, with all three successfully demonstrating that they can viably support a quantum computer. The real question is, which will find the backing and the best use-case first?